Prediction Accuracy as a Percentage

I realize this is simpler than most of the stuff here, but I would really appreciate any help with this.

I'm trying to get a percentage of accuracy for column A.
Column A is the predicted value (forecast), Column B is the actual value (the result), so how accurate was the Column A prediction?

For example, two scenarios: let's say the forecast was 18 but the actual value was 15 (forecast higher). And the same forecast model also gave a forecast of 35 on another object, but the actual value was 40 (forecast lower). I want to combine the two forecasts of the same model and give the accuracy as a percentage out of 100%. Can anybody help me to come up with a single formula for this?

I'm not a statistics guy - I've tried to google this but to no avail. Any help much appreciated!

That makes sense.

In the first case it will give you 0,8 aka 80%, which means the prediction failed by 20% since the actual value come to only 15.
In the second case it will give you 1,13 aka 113% which means the prediction surpassed the prediction by 13% since the actual value come to 40 instead of 35.

If the prediction was 20 and the actual result 20 it will give you 1 - which means exact predicion.

Nice feature. I can use this by myself.

Yeah it's like FCarv said.

The formula works perfectly. I got close with my rudimentary knowledge of math but I didn't know about the ABT function. Thank you very much, this is very useful.